JP2001275098A - Server system and data transfer method - Google Patents

Abstract

(57) [Problem] To provide a simultaneous multiple access control technique in which a large number of clients can be connected in a client-pull video-on-demand server system. In a server system, a client pull video stream data is extracted from an information storage unit to a temporary storage unit and transferred to a client.
When the data is read from the information storage unit 15 to the temporary storage unit 14, the number of clusters of video stream data to be acquired at which time is a future video stream data read plan based on the current read state. A data read plan unit 22 for creating a read plan line, and input of video stream data from the information storage unit 15 to the temporary storage unit 14 and data transfer from the temporary storage unit 14 to the client according to the created read plan line. Input / output control means 21 for controlling output.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a server system and a data transfer method capable of connecting a large number of servers by reducing the load by reducing the frequency of access to an external storage device. Related to simultaneous multiple access control technology.

[0002]

2. Description of the Related Art Distributing different information selected and requested by individual clients to clients in response to the request.
In a so-called video-on-demand system, a center facility called a video-on-demand server that stores and manages video information and related data and sends out the video data and related data according to a client's request is provided. The video-on-demand server is usually composed of one or a plurality of workstations or personal computers, and the number thereof is configured according to the size of the video-on-demand system.

[0003] As shown in FIG. 8, the video-on-demand server 50 includes a management server 51, an application server 52, a stream server 53, and an input server 54. Among them, the management server 5
Reference numeral 1 denotes management of the entire server system (management of address information on the hard disk 55 where applications and video programs are stored, or management of billing information for each client). It is responsible.

[0004] The application server 52 is a server that provides various interactive applications. That is, information is frequently exchanged with each client, and therefore, high-speed responsiveness that can withstand this exchange is required.

[0005] The stream server 53 is a server for storing and transferring video streams, and is connected to a communication device 56 for transferring data to clients. It is also connected to a hard disk 55 for storing video data and audio data. Therefore, the stream server 53 is required to be able to continuously supply stream data from the hard disk 55 at a stable speed.

[0006] The input server 54 includes a management server 51.
It plays a role in the input / output control device 57 in the same manner as that described above, but its role is to control the input / registration of video contents. A video-on-demand server system is configured by these subsystems such as several servers.

Here, there is a client pull (client-led) type in the video on demand system. In this method, in response to a request (control) for providing information from a client, a video stream corresponding to the request is transmitted to the client, and the client makes a request for information. The information provider does not provide one-sided information to the client,
It is provided to clients on demand. Note that requests from these clients can be specified in a home page described in HTML language in units of stream content titles. When a request is made for each packet, which is a transmission unit of a stream, the request can be made by the HTTP protocol.

Here, in the conventional client-pull video-on-demand system, when there is a request for information provision from a client, information of a storage device such as the hard disk 55 is acquired for each cluster through a hard disk cache. . In other words, whenever there is a request for information acquisition from a different client, in general, a hard disk access is performed each time, and the acquired data is stored in a hard disk cache.
The data is obtained from the hard disk cache without accessing the hard disk.

In addition, the method of automatically acquiring not only the data of the cluster unit but also the data of the cluster unit of a plurality of destinations which is assumed to be related to the specific cluster in the form of the hard disk cache, ie, read Even in the case of the ahead cache method, in response to a request for information acquisition from a client, a specific cluster and information of a cluster unit related thereto are acquired each time.

[0010]

By the way, even when the above-mentioned hard disk cache is used, if the cache hit rate is low, the hard disk is accessed in response to a data acquisition request from a client. There are situations where access occurs frequently. A typical example is a case where a large number of clients request different stream titles and are simultaneously connected. Here, the hard disk access time is calculated as follows.

Time required for moving the head in the radial direction of the magnetic disk to access a predetermined position = Average rotation waiting time (time for moving by 1/2 cycle) + average seek time (time for moving by 1/3 radius)

From this, it can be seen that the cumulative time of the head operation by the actuator in the hard disk 55 and the cumulative time of the magnetic disk rotation waiting become longer.
If the data acquisition is not performed by the predetermined time due to the increase in the access time, the continuity and punctuality of the stream reproduction on the client is lost, and the time is over.

Then, the stream server 53
However, the connection of the client is restricted.

In this respect, a similar problem occurs in the read-ahead cache system. In other words, even if data for a plurality of clusters, which is assumed to have a relationship with a specific cluster, is automatically acquired, the acquisition is not managed by a server or the like. Therefore, when a plurality of clients are connected, the waiting time is accumulated and it takes a very long time to acquire data.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a client-pull video-on-demand server system.
An object of the present invention is to provide a simultaneous multiple access control technique capable of connecting a large number of clients.

[0016]

In order to achieve the above-mentioned object, the present invention provides a server system for extracting client-pull video stream data from information storage means to temporary storage means and transferring the video stream data to a client. When data is read from the information storage unit to the temporary storage unit at the request of the client, the number of clusters of video stream data to be obtained at which time is determined based on the current video stream data reading condition. A data reading plan means for creating a reading plan line, which is a reading plan of video stream data, and video stream data actually input from the information storage means to the temporary storage means according to the reading plan line created by the data reading planning means. And output control means for controlling the output of data to the client from the temporary storage means and has a.

Thus, the data reading planning means
In this configuration, a reading schedule line is created and the reading of the video stream data is managed so as to follow this. Therefore, if the video stream data is acquired while managing the reading of the video stream data so as to follow the read plan line, the number of accesses to the information storage means can be reduced and the video stream data can be extracted to the temporary storage means. Thus, if the number of accesses to the information storage means is reduced,
As a result, the time spent in the information storage means can be reduced, and as a result, many clients can be connected with limited hardware resources.

In another invention, in addition to the above-mentioned invention, the temporary storage means is a cache memory.

Therefore, the processing of the video stream data can be performed at high speed.

In another aspect of the present invention, in addition to the above-mentioned inventions, the data reading planning means is further connected to a determining means for determining the data type of the video stream data. The data reading planning means determines whether or not to create a reading plan line based on the determination result of the type of stream data.

For this reason, according to the type of video stream data, a read plan line is created only for video stream data of an appropriate type. Therefore, there is no waste in creating the read planning line.

In another aspect of the present invention, the data reading determining means includes, in a cluster unit of adjacent video stream data, a reading start time point of the first data to be read first in a first cluster from which the data is read first; The inclination of the read planning line is set to be equal to or greater than the inclination of a line connecting the read end time of the data read last in the second cluster in which data is read last.

By setting the reading of the video stream data as described above, it becomes possible to supply the video stream data while sufficiently satisfying the video on demand required by a specific client. In addition, by setting the inclination of the read planning line in a specific client in this way, if this is followed, it becomes possible to spend time reading video stream data in another client. As a result, a large number of clients can be connected with limited hardware resources.

Further, according to another invention, in addition to the above-mentioned inventions, the temporary storage means further comprises a first temporary storage for extracting and storing video stream data from the information storage means for each pack in a first-in first-out manner. Means for transmitting the video stream data stored in the first temporary storage means to the client or outputting the video stream data from the first temporary storage means in a first-in first-out manner; Second temporary storage means for storing only the data files in the pack.

For this reason, in the first temporary storage means, it is possible to successively extract and store video stream data from the information storage means. In addition, this
Of the video stream data extracted from the first temporary storage means, only the data files in the data pack are stored in the second temporary storage means, so that only the necessary parts except the blank part in the data pack are stored. be able to. Then, if only a necessary portion of the video stream data is called at the time of transfer from the second temporary storage means to the client, high-speed processing can be performed.

According to another aspect of the present invention, there is provided a data transfer method for transferring video stream data stored in an information storage unit of a server system to a client on a video-on-demand basis. When temporarily storing the stored video stream data in the temporary storage means, the number of clusters of video stream data to be obtained at which time is determined based on the current video stream data reading condition. A read plan line creating step for creating a read plan line, which is a read plan of the video stream data, and video stream data actually stored in the information storage means in accordance with the read plan line created in the read plan line creating step. Read above And reading step of storing in 憶 means is for anda transfer step of transferring the stored video stream data to the client in the temporary storage by reading process.

As described above, according to the present invention, in the data read plan line forming step, a read plan line is prepared and the reading of the video stream data is managed so as to follow the read plan line. Therefore, to follow this load plan line,
If the video stream data is acquired while managing the reading of the video stream data in the reading step, the number of accesses to the information storage means can be reduced and the video stream data can be extracted to the temporary storage means. Further, as described above, if the number of accesses to the information storage unit is reduced, the time spent in the information storage unit can be reduced accordingly, and as a result, many clients can be connected with limited hardware resources. be able to.

Further, in another invention, in addition to the above-described invention, a discriminating step for discriminating the data type of the video stream data is provided prior to the reading plan line creating step. A determination is made as to whether or not to create a read plan line in the data read planning step based on the determination result of the type of stream data.

For this reason, a read plan line is created only for video stream data of an appropriate type according to the type of video stream data. As a result, there is no waste in creating the read planning line.

Further, in another invention, the temporary storage means is the first storage means.
In the reading step, the video stream data is extracted from the information storage means for each pack in a first-in first-out manner and stored in the first temporary storage means. 1 when the video stream data stored in the first temporary storage means is transferred to the client or in the first-in first-out manner.
And a second reading step of storing only the data file in the data pack of the video stream data in the second temporary storage means when the video file is output from the temporary storage means.

Therefore, in the first reading step, video stream data can be successively extracted from the information storage means and stored in the first temporary storage means.
Also, of the video stream data read in the first reading step, only the data files in the data pack are stored in the second temporary storage means in the second reading step. Only necessary parts can be stored. And
High-speed processing can be performed by calling only a necessary portion of the video stream data at the time of transfer from the second temporary storage means to the client.

According to another aspect of the present invention, in addition to the above-mentioned inventions, in the reading step, when the inclination of the reading plan line created in the reading planning line creation step is smaller than a reference value, the reading planning line Regardless of this, the video stream data is read.

As described above, when the inclination of the read planning line is smaller than the reference value, even if the video stream data is read without any management, the transfer of the video stream data to the client may be hindered. Absent. Not only that, but not managing in such cases,
Processing can be simplified, and the speed of data transfer to the client can be improved.

Further, according to another invention, in addition to the above-mentioned inventions, in the reading step, if the time interval for actually reading the video stream data is longer than the reference value, regardless of the reading plan line, It reads video stream data.

As described above, when the time interval for reading the video stream data is longer than the reference value, the transfer of the video stream data to the client is not affected even if the reading of the video stream data is not performed at all. Never come out. In other words, in this case, since the waiting time for transferring from the temporary storage means to the client is long,
Reading time from the information storage means to the temporary storage means does not matter. Therefore, by not managing in such a case, the processing can be simplified, and the speed of data transfer to the client can be improved.

[0036]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIGS.

FIG. 1 is a diagram showing the configuration of a video-on-demand server system 1 according to the present invention. In this figure,
The video-on-demand server system 1 has a video-on-demand server 2. The video-on-demand server 2 functionally includes a management server 3, an application server 4, a stream server 5, and an input server 6.

And, among these, the stream server 5
Is connected to a hard disk 15 as information storage means for storing video data and the like. Also, the management server 3
The input server 6 plays a role as the input / output control device 21. In particular, the management server 3 manages a URL (Uniform Resource Locator) and client identification information when the client accesses the video-on-demand server 2. Further, the stream server 5 is connected to the communication device 20, and can execute various interactive applications with the client.

The management server 3, the application server 4, the stream server 5, and the input server 6 are connected to each other via a communication path such as a LAN in a large-scale configuration, so that signals and data can be transmitted to each other. In such a large-scale configuration, each of these servers 3
6 to 6 are workstations and personal computers by themselves. However, in a medium-scale or small-scale configuration instead of a large-scale configuration, a single or a small number of workstations or personal computers are configured to also perform the above-described functions. The following description is based on a single workstation or personal computer.

FIG. 2 is a diagram showing the configuration of a small-scale video-on-demand server 10 composed of a single personal computer. The video-on-demand server 10 has a CPU 11 that controls the entire server. This CP
U11 fulfills the functions of various servers in the large-scale video-on-demand system described above, together with other components of the personal computer.

The CPU 11 is connected to the program memory 12. As a result, the application program read into the program memory 12 is executed to enable data arithmetic processing.

This CPU 11 is connected to the cache allocation table 13. The cache allocation table 13 links a location on a file stored on the hard disk 15 as an information storage unit with a location on the cache 14.

A cache 14 as a temporary storage is connected to the CPU 11 and the cache allocation table 13. This cache 14 has an A cache 14a as a second storage unit.
And a B cache 14b as a first storage unit. The cache 14 can exchange data with the CPU 11 at a higher speed than the program memory 12.

The CPU 11, the cache allocation table 13 and the cache 14 are connected to a hard disk 15 which is an external storage device. The hard disk 15 stores video data and music data requested by the client.

Here, the configuration of the hard disk 15 is shown in FIG. The hard disk 15 has a plurality of magnetic disks 31 that record video data and music data. A head 32 for taking out data is arranged opposite to the magnetic disk 31. Also, head 3
Numeral 2 is connected to an actuator 33 so that the head 32 can be driven in the radial direction of the magnetic disk 31. Further, the magnetic disk 31 is attached to a magnetic disk rotation motor 34. Thereby, the magnetic disk 31 is driven to rotate. The head 32 is driven by the actuator 33 and the magnetic disk rotation motor 34, so that information recorded at a specific position on the magnetic disk 31 can be read.

Here, usually, the CPU 11, the cache allocation table 13, the cache 14, and the hard disk 15 constitute a single server such as a personal computer or a workstation. However, the server may be constituted by a plurality of personal computers and workstations instead of a single personal computer and workstation.

The CPU 11, the cache allocation table 13, and the cache 14 are connected to a network I / F 16 constituting the communication device 20. Each client connects to the video-on-demand server 10 via the network I / F 16.

FIG. 4 shows a configuration in which the video-on-demand server 10 is functionally grasped. Explaining based on this figure, each client is connected via a communication device 20 provided in the video-on-demand server 10. The communication device 20 is connected to an input / output control device 21 as input / output control means. The input / output control device 21 corresponds to the management server 3 and the input server 6 described above. In the configuration of FIG. 2, the CPU 11 and the program memory 12 and the application programs and the like in the program memory 12 correspond. is there.

The input / output control device 21 has a hard disk read scheduler 2 as a data read planning means.
2 (hereinafter abbreviated as HDRDSC). This is because the input / output control device 21
And controls how much of the video data is stored in the cache 14. This makes it possible to efficiently deliver a video stream based on video requested by a plurality of clients to each client.

That is, the HDRDSC 22 uses the head 32 of the hard disk 15 to (1) when (2) video stream data requested by any client,
(3) A control command for how much data is to be read is output to the input / output control device 21.

Here, the video stream data is composed of moving image data (including audio data). As shown in FIG. 5, the moving image data basically includes
Moving image data recognition code (code for identifying the data format), data specifying the next data acquisition range, accompanying data covering the entire moving image data, data required for each channel, and packet index delimited for each packet And the packet code are stored in this order.

The packet constituting the packet data is a group of data divided in a certain unit.
In the present invention, a series of data acquisition accompanying stream playback,
Represents a data acquisition unit that is synchronized with the processing cycle. Specifically, the packet includes a packet index including three sections of a packet header, data acquisition information, and data reproduction information, and a packet code including a reproduction data section.

The four sections will be further described with reference to FIG. 5. The packet header indicates a code for recognizing the beginning of a packet, a pointer to the next packet, a pointer to a past packet, and a packet configuration. Start and end pointers are included. The data acquisition information includes a data storage start position, a data storage end position, and the like. The information for data reproduction includes basic items for reproduction such as the number of frames present in a packet, an appropriate time to start reproduction processing of the packet, reproduction time information in frame units, frame position information, and frame unit information. It mainly consists of scalability hierarchy information. What is playback data?
This refers to audio data or image data to be reproduced.

For example, a video stream is configured with one packet composed of a plurality of sector units as one cluster unit.

In the video-on-demand server 10 having the above-described configuration, when there is a data acquisition request from a client, a video stream data transmission technique for transmitting a video stream to the client is a simultaneous multiple access control technique. The data transfer method will be described below with reference to the drawings.

First, the client requests the video on demand server 10 to send a video stream of specific video data. In this case, the request from the client is transmitted to the communication device 20 including the network I / F 16.
Is input to the input / output control device 21 via the. Since there are multiple clients, at a specific time,
This is a state in which a large number of clients are requesting acquisition of separate or identical video stream data.

Then, in the input / output control device 21 to which the request from the client is first input, the HDRDSC 22
For each client, a signal indicating that a request to acquire specific video stream data has been made is transmitted. Then, the input / output control device 21
Receives an instruction to acquire video stream data. This acquisition schedule command is a schedule for acquiring video stream data from the hard disk 15 to the B cache 14b. Thereafter, acquisition of video stream data from the hard disk 15 is performed by the input / output control device 21 based on a given data acquisition instruction.

This situation will be described with reference to FIG. This figure shows a state in which three clients A, B and C are connected. Then, inside the hard disk 15, as shown in FIG. 3, the head 32 of the hard disk 15 is driven by an actuator 33 and a magnetic disk rotation motor 34. This allows the head 32 to access a specific area of the magnetic disk 31 on which the video stream data is stored, and obtain desired video stream data. Then, the video stream data is read into the B cache 14b.

That is, in FIG. 6 described above, a portion indicated by a black line in one or a plurality of clusters (indicated by reference numeral 30 in FIG. 6) corresponding to the video stream is a portion from the head 32 to the B cache 14b. Indicates the time when the video stream data is actually read. A white portion other than the black line in each cluster 30 indicates a time during which the head 32 accesses a specific area of the magnetic disk 31 where the video stream data is stored or a time when the head 32 is stopped. The time during which the video stream data is actually read from the head 32 to the B cache 14b, which is the portion indicated by the black line, is instantaneous, and is therefore represented by a thin black line in this figure. However, the time during which the head 32 moves is overwhelmingly longer than the reading time, and is therefore represented by a white portion in this figure.

When the head 32 accesses the area of the magnetic disk 31 where the video stream data is stored, the magnetic disk 31 is rotated at a high speed. Shorter than the travel time in the direction. Here, client A at a specific time
Is required to be in the same rotation direction or in the vicinity of this rotation direction.
In addition, in the video stream data constituting one story, a data transfer procedure until the story ends is predetermined.

For this reason, the video stream data requested by the client A is acquired for a plurality of clusters which sufficiently satisfies the transfer time limit as long as the data transfer to another client is not delayed. Then head 3
Even if the user 2 goes to obtain video stream data for another client existing in another area, it is possible to respond to the request of the client A for the time being.

After the data acquisition for a plurality of clusters of the client A is completed, the actuator 3
3 and the magnetic disk rotation motor 34 are operated to move the head 32 to an area of the magnetic disk 31 where data required by the client B is stored. After this,
The same operation as that performed by the client A described above is also performed by the client B. The same applies to the client C.

Here, a more detailed description will be given of the case where the video stream data requested by the client A, one of the three, is acquired. In this case, in consideration of the relationship with the data acquisition requests of the other clients B and C, how long the client A should acquire the data and how much data should be acquired are considered. That is,
Since one head 32 can read only one specific data at a time,
The HDRDSC 22 determines how many cluster units of data are acquired at which time. Also in this case, it is determined in consideration to sufficiently satisfy the transfer time limit of the video stream data to the client A.

The source of the decision is as follows:
FIG. 8 is a reading plan diagram as shown in FIG. 7. This read plan line indicates a plan for acquiring video stream data from the hard disk 15 to the B cache 14b in the specific client A. In this figure, the horizontal axis indicates the data reading time from the hard disk 15 in the B cache 14b, and the vertical axis indicates the number of bytes of the data (file) to be read.

When data is actually read by the head 32, first, video stream data requested by the client A is obtained for two clusters enclosed by broken lines. Then, by reading the video stream data for the two clusters, the HDRDSC 22 creates a read plan line.

The first read data (file) of the first cluster from which the data was read first
At the beginning of reading and the second
The inclination of the read planning line is set to be equal to or greater than the inclination of the line connecting the end of reading of the data (file) most recently read in the cluster. Then, the subsequent reading of the video stream data from the hard disk 15 is set so as to follow the reading plan line (the one that plans to complete the reading earlier than the actual reading result).

In this case, when video stream data is supplied to the client A, data transfer can be always performed with a certain margin.

Here, in FIG. 7, the fact that the head 32 actually reads data is shown as a black line, and in this figure, three data (files) form one data pack. Has become. One cluster is formed by collecting these data packs. Note that a cluster may be configured from one data pack as in the other cases in the present invention.

In the example shown in FIG. 7 described above, the read plan line is created by acquiring the first two clusters, but the read result immediately after the transfer is read according to the read state of the video stream data. The reading schedule line may be created at any time based on. By doing so, it is possible to transfer the video stream data while correcting the read plan line as needed.

However, when reading video stream data from the hard disk 15, the calculated inclination of the read planning line may be smaller than the set reference value. In this case, the transfer time of the video stream data in the client A is short, but if it is smaller than such a reference value, HDRD
Even if the reading of the video stream data from the hard disk 15 is not managed by the SC 22, the video on demand for the client A is inevitably satisfied. This is based on a small transfer amount of video stream data requested by the client. Therefore, in this case, it is possible to adopt a configuration in which the HDDRSC 22 does not manage reading of the video stream data at all.

Even when the time interval for reading the video stream data from the hard disk 15 is longer than a certain reference value, even if the HDRDSC 22 does not manage the reading of the video stream data from the hard disk 15, the client A Video on demand. Again, HDRDS
It is possible to adopt a configuration in which reading of video stream data is not managed at all in C22.

Further, in the case where the HDRDSC 22 creates a read planning line, a configuration may be provided in which a determination means for separately determining the type of video stream data is provided. The determining means determines whether or not the video stream data is of a type for which a read plan line needs to be created, based on the extension of the video stream data. Then, based on this determination, a reading plan line is created only for the video stream data that needs to be created.

Here, a more detailed difference from the case where the read-ahead cache is used will be described. In the read-ahead cache, when data of a specific cluster is being read, data of a plurality of clusters that are related to the data of the specific cluster are also read together. In this regard, it can be said that the present invention and the case where the read-ahead cache is used have similar surfaces.

However, when a read-ahead cache is used, if many clients are connected at the same time, the time for reading data for many other clients becomes longer. As a result, the transfer of the video data stream data of a specific client is eventually stopped, and a situation occurs in which the client cannot obtain desired video stream data.

In order to prevent such a stop of the video stream, the number of client accesses may be limited. However, if the number of accesses is restricted in such a manner, a situation occurs in which it is impossible to meet the demands of a large number of clients who desire access.

For this reason, in the present invention, instead of using a read-ahead cache that does not perform any control in the video-on-demand server 10 to acquire data, the HDRDSC 22 is used to create a read plan line as a data acquisition schedule. It is created and the data is obtained from the hard disk 15 based on this. Thus, the configuration is such that a greater number of clients can be accessed while having limited hardware resources.

When necessary data is read by the B cache 14b, only the data file of the data pack of the video stream data requested to be obtained by the client is copied from the B cache 14b to the A cache 14a. . And B cache 1
In 4b, reading of new video stream data is sequentially performed for each data pack by a first-in first-out method (abbreviated to fifo; first in first out). That is, by copying only the necessary data requested to acquire the video stream data to the A cache 14a, the maximum data can be read for the limited capacity of the cache 14,
It is possible to transfer to the client.

Then, the video stream data copied to the A cache 14a is transferred from here to the desired client via the communication device 20. Thereby, the client can obtain desired video stream data.

The above procedure is performed simultaneously for information terminals such as personal computers of many other clients. Then, each client can browse each program including desired video data and audio data on demand.

According to the video-on-demand server system 1 and the data transfer method having such a configuration, the HDRDS
C22 is provided, and a read planning line is created by the HDRDSC22. Then, the video stream data is acquired from the hard disk 15 by the input / output control device 21 so as to follow the created read plan line. As described above, if the video stream data is obtained while being managed, the number of accesses to the hard disk 15 can be reduced, and the video stream data can be extracted to the cache 14.

As described above, if the number of accesses to the hard disk 15 is reduced, the hard disk 15
The time lost for accessing a specific area in which data is stored can be reduced, and as a result, many clients can be connected with limited hardware resources.

When a read planning line is created by the HDRDSC 22, the read start time of the first read data (file) of the first cluster of the first read data and the second read second data of the last read data The inclination of the read planning line is set to be equal to or greater than the inclination of the line connecting the end of reading the data (file) most recently read in the cluster.

For this reason, if the subsequent reading of the video stream data is obtained in accordance with the read plan line, the data transfer can be performed while satisfying the video on demand for the client.

Further, the cache 14 is stored in the A cache 1
4a and two B caches 14b. Accordingly, when the video stream data is obtained from the hard disk 15, not only can the video stream data be exchanged with the CPU 11 at a high speed, but also the necessary video stream data except for the blank portion in the data pack can be obtained. Only part is B cache 14
b can be stored in the A cache 14a.
If only a necessary part of the video stream data is called from the A cache 14a, processing can be performed at a higher speed.

In the B cache 14b, video stream data is extracted from the hard disk 15 for each data pack in a first-in first-out manner and sequentially stored. Further, only the data requested to be obtained by the client is copied from the B cache 14b to the A cache 14a. For this reason,
The B cache 14b reads the maximum amount of video stream data for a limited capacity, while the A cache 14a can promptly transfer the data in response to a data acquisition request from a client.

If the next acquisition request exists in the B cache, the data is transferred from the A cache 14a at high speed only by copying the data from the B cache 14b without accessing the hard disk 15. It becomes possible.

Further, in the configuration including the discriminating means for discriminating the type of video stream data, the type of video stream data is discriminated by an extension or a data header, and an appropriate type of video stream is determined according to the type of data. Create a read plan line for stream data only. Therefore, there is no need to create a loading feature line,
For video stream data, there is no need to uselessly create a read planning line, and thus, by eliminating waste, it is possible to improve overall processing efficiency.

When reading the video stream data, if the calculated inclination of the read plan line is smaller than the set reference value, the HDRDSC 22 does not manage the reading of the video stream data from the hard disk 15. You can also. By providing such a reference value, it is possible to easily determine whether or not the management of the reading of the video stream data is necessary in the HDRDSC 22, and if the inclination of the reading plan line is smaller than the reference value, it is assumed that the reading is not managed at all. If this is the case, the processing procedure is simplified and the overall processing efficiency is improved.

Further, it is possible to adopt a configuration in which the reading of the video stream data from the hard disk 15 is not managed by the HDRDSC 22 when the time interval for reading the video stream data from the hard disk 15 is longer than a certain reference value. Even when such a reference value is provided, it is possible to easily determine whether or not the management of the reading of the video stream data is necessary in the HDRDSC 22, and if the reading is not managed at all when the time is longer than the reference value, the processing is performed. The procedure is simplified, and the overall processing efficiency is improved.

Further, according to the present invention, as described below, there is an advantage that occurs when an individual client is captured, and an advantage that occurs when a large number of clients are captured collectively. That is, it can be said that there are merits over two aspects. Both viewpoints are common in that the total access time of the head 32 to the area where the data of the magnetic disk 31 in the hard disk 15 is stored can be reduced.

First, each client reads video stream data from the hard disk 15 within the range of the read planning line. As a result, the transfer of the video stream data is not interrupted, and a video stream of stable quality is supplied. Therefore, a video-on-demand system closer to the ideal of the client is obtained.

Although the number of accessible clients is limited within a certain range due to hardware restrictions such as the access time of the head 32, the number of accessible clients can be increased. Thereby, it is possible to supply a video stream to a larger number of clients on demand with limited hardware resources.

Although the embodiment of the present invention has been described above, the present invention can be variously modified. Hereinafter, this will be described.

In the above embodiment, a video on demand server system integrated in one place is assumed.
Of course, it is also possible to apply the present invention to a local server for decentralizing access. Providing a local server as described above makes it possible to connect a larger number of clients.

Also, as a video on demand server,
In the above-described embodiment, the case of accessing moving image data including video data and audio data, for example, the case of browsing each program, has been described. Is applicable.
In this case, the information terminal of the client may be not only a personal computer but also a home game device. Further, the present invention may be applied not only to a personal computer but also to a portable information terminal.

[0096]

According to the present invention, the data reading planning means creates a reading planning line and manages reading of the video stream data in accordance with the drawing planning line. Therefore, if the video stream data is acquired while managing the reading of the video stream data so as to follow the read plan line, the number of accesses to the information storage means can be reduced and the video stream data can be extracted to the temporary storage means. As described above, if the number of accesses to the information storage unit is reduced, the time spent in the information storage unit can be reduced accordingly, and as a result, many clients can be connected with limited hardware resources. it can.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a video-on-demand server system according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a small-scale video-on-demand server in the video-on-demand server system of FIG. 1;

FIG. 3 is a diagram showing a configuration of a hard disk in the video-on-demand server of FIG. 2;

FIG. 4 is a diagram showing a configuration in which the video on demand server of FIG. 2 is functionally grasped.

FIG. 5 is a diagram showing details of moving image data in video stream data.

FIG. 6 shows the video on demand server of FIG.
It is a figure showing the state where a human client is connected.

FIG. 7 shows the HDRDS of the video on demand server of FIG.
9 shows a read planning line created in C.

FIG. 8 is a diagram showing a configuration of a conventional video-on-demand server.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Video on demand server system 2, 10 ... Video on demand server 3 ... Management server 4 ... Application server 5 ... Stream server 14 ... Cache (temporary storage means) 14a ... A cache (second temporary storage means) 14b ... B Cache (first temporary storage means) 15 ... Hard disk (information storage means) 20 ... Communication device 21 ... Input / output control device (input / output control means) 22 ... Hard disk read scheduler (data read determination means) 32 ... Head

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 13/00354 G06F 13 / 00354D

Claims (10)

[Claims] 1. A server system for extracting video stream data of a client pull from an information storage means to a temporary storage means and transferring the video stream data to a client, wherein the data is transferred from the information storage means to the temporary storage means at the request of the client. Data for creating a read plan line, which is a read plan for future video stream data, based on the current video stream data read status, regarding how many cluster units of video stream data to acquire at which time to read. In accordance with a read plan line created by the read plan means and the data read plan means, video stream data is actually input from the information storage means to the temporary storage means and data is transferred from the temporary storage means to the client. Server system, characterized in that a, input and output control means for controlling the output of. 2. The server system according to claim 1, wherein said temporary storage means is a cache memory. 3. The data reading plan means is connected to a judging means for judging a data type of the video stream data, and the data reading planning means is judged based on a judgment result of the type of the video stream data by the judging means. 3. The server system according to claim 1, wherein a determination is made by the means as to whether or not to create a read planning line. 4. The data reading determining means, in cluster units of adjacent video stream data, start reading of the first read data of a first cluster from which data is read first, and end the data of the first cluster. 4. The method according to claim 1, wherein the inclination of the read planning line is set to be equal to or larger than the inclination of a line connecting the read end time of the data read last in the second cluster to be read. Server system as described. 5. The temporary storage means includes: first temporary storage means for extracting and storing video stream data from the information storage means for each pack in a first-in first-out manner; and storing in the first temporary storage means. When transferring the video stream data to the client or when outputting the video stream data from the first temporary storage means in a first-in first-out manner, only the data file in the data pack of the video stream data is stored. The server system according to any one of claims 1 to 4, further comprising two temporary storage units. 6. A data transfer method for transferring video stream data stored in an information storage means of a server system to a client on a video-on-demand basis, wherein the video stream data is stored in the information storage means requested by the client. When the video stream data is temporarily stored in the temporary storage means, the number of cluster units of video stream data to be obtained at which time is determined based on the current video stream data reading condition. A read plan line creating step of creating a read plan line, which is a read plan, and reading video stream data actually stored in the information storage means in accordance with the read plan line created in the read plan line creating step. Above temporary And reading step of storing in 憶 means, data transfer method, characterized by comprising a transfer step of transferring the video stream data stored in the temporary storage means to the client by the reading process. 7. A discriminating step for discriminating a data type of the video stream data is provided prior to the reading plan line creating step, and the data reading is performed based on a discrimination result of the video stream data type in the discriminating step. 7. The data transfer method according to claim 6, wherein it is determined whether or not to create a read planning line in the planning process. 8. The temporary storage means comprises a first temporary storage means and a second temporary storage means, and in the reading step, video stream data is read from the information storage means in a first-in first-out manner. A first reading step of extracting each pack and storing it in the first temporary storage means, and when transferring the video stream data stored in the first temporary storage means to the client or by a first-in first-out method A second reading step of storing, in the second temporary storage means, only the data file in the data pack of the video stream data when output from the first temporary storage means. The data transfer method according to claim 6. 9. The data transfer method according to claim 6, wherein in the reading step, a slope of the read planning line created in the reading plan line creating step is smaller than a reference value. A method of reading video stream data irrespective of a read plan line. 10. The data transfer method according to claim 6, wherein in the reading step, when a time interval for actually reading the video stream data is longer than a reference value, A data transfer method characterized by reading video stream data regardless of a read planning line.